Not sure how much I trust the model to not use hindsight, but it promised that it did.
Also fucked up some of the years. Couldn’t be asked to spend more than 5 minutes. But with some actual prompt engineering, Gemini probably would do better than Duke
For funsies I asked chatgot: “Based on information known on each day of the 2025 draft, what players would you have selected for the Cincinnati Bengals?”
6 comments
Lol, I approve.
Chat GPT would just draft whatever players people on Reddit are saying we should draft.
https://preview.redd.it/rgzwqjvxc35g1.jpeg?width=1320&format=pjpg&auto=webp&s=fc4617ac0e663fd8cc6dbea18d56e07629c036e6
Not sure how much I trust the model to not use hindsight, but it promised that it did.
Also fucked up some of the years. Couldn’t be asked to spend more than 5 minutes. But with some actual prompt engineering, Gemini probably would do better than Duke
For funsies I asked chatgot: “Based on information known on each day of the 2025 draft, what players would you have selected for the Cincinnati Bengals?”
Reply minus fluff:
Round 1 (Pick 17)
Pick: Malaki Starks — S, Georgia
Round 2 (about pick 49)
Pick: Walter Nolen — DT, Ole Miss
Round 3 (early/mid third)
Pick: Xavier Watts — S, Notre Dame
Pick: Jalon Walker — EDGE / OLB (versatile pass rusher / hybrid LB-EDGE)
Round 5
Pick: Interior O-Line — maybe a guard or center, depending on availability
Round 6 / 7
Pick: Depth RB or developmental DL / edge perhaps
If you actually ask chat gpt to draft 90% it will pick players that have already been drafted
Katie Blackburn after firing Duke Tobin: “Okay ChatGPT, what should our first move be?”
ChatGPT: “You should hire Duke Tobin”